Ca2+ is the most ubiquitous signaling molecule in the human body, regulating numerous biological functions that include heart beat, muscle contraction, neural function, cell development, and proliferation, by fluxing between the subcellular compartments with different amplitudes and durations[1]. The membrane-based organelle endo/sarcoplasmic reticulum (ER/SR) lumen, which occupies less than 10% of cell volume, stores more than 90% of intracellular Ca2+ and is pivotal in controlling Ca2+ signaling. It can produce intrinsic Ca2+ release and propagation of Ca2+ oscillations[2-4]. Ca2+-mobilization agonists such as ATP, ionomycin, histamine, and glutamine will activate Ca2+ receptors and pumps, such as inositol 1,4,5-trisphosphate receptor (IP3R), to release Ca2+ from the ER into the cytosol[5-7], which results in a rapid decrease of ER Ca2+ (from mM at the resting state to μM in excited state). The removal of these agonists will help Ca2+ refill the ER through membrane channels such as sarco(endo)plasmic reticulum Ca2+-ATPase (SERCA). The alternation of Ca2+ concentration activates various intracellular Ca2+ sensing (trigger) proteins, such as calmodulin (CaM), troponin C (TnC) and other ion channels, through conformational changes that occur upon binding to Ca2+[8]. These activated Ca2+-sensor receptors will further regulate numerous cellular processes and events. Recent studies indicate that Ca2+ signaling is important for homeostatic handling of cardiovascular functions[9-11]. In cardiomyocytes, cardiac relaxation and contraction is regulated by the periodic change of intracellular Ca2+ concentration and the proteins associated with the sarcoplasmic reticulum (SR), a homologue of ER[12, 13]. The cardiac ryanodine receptor (RyR2), inositol (1,4,5)-trisphosphate receptor (IP3R) and the sarcoplasmic reticulum Ca2+-ATPase 2a (SERCA2a) are three pivotal portals for the Ca2+ mobilization during this agonist-induced process. Heart failure caused by dysfunction of these two proteins, associated with abnormal Ca2+ handling, is becoming increasingly evident in data collected both from animals and humans[14-17]. A Ca2+ indicator to monitor ER/SR Ca2+ concentrations with fast release kinetics, and the capability to quantitatively detect Ca2+ signaling in specific subcellular organelles will have a significant impact on the understanding of the molecular basis of Ca2+ signaling and homeostasis in cardiac development and diseases.
The initial measure of ER Ca2+ dynamics was achieved using the Ca2+ dye Mag-fura-2 in plasma membrane-permeabilized live cells. In contrast to Ca2+ dyes, fluorescent protein (FP)-based Ca2+ indicators with genetically encoded chromophores can detect Ca2+ signaling in subcellular organelles with high spatial and temporal resolution. They consist of a Ca2+-modulated protein, either calmodulin or troponin C, coupled to a single fluorescent protein to generate sensors, such as GCaMP (11), or dual fluorescent proteins, such as Cameleon. Modifying Cameleon at its Ca2+ binding loops or CaM's peptide-interaction surface generated several ER/SR sensors, which have been applied to excitable cells with some limitations. Directly monitoring fast ER/SR Ca2+ dynamics in excitable cells is still new territory.
As a secondary messenger, calcium ions regulate many biological processes in various intracellular compartments through interactions with proteins. Calcium is involved in muscle contraction (including heartbeat), vision, and neuronal signaling. Calcium binding proteins exhibit different calcium binding affinities with Kd ranging from 0.1 μM to mM, which are essential for their responses to various stimuli through the temporal and spatial changes of calcium and calcium homeostasis. For example, extracellular calcium-modulated proteins with multiple calcium binding sites, such as cadherins and calcium-sensing receptors, have dissociation constants in the submillimolar to millimolar range. Calsequestrin, a major calcium binding protein in the endoplasmic reticulum (ER), has a relatively weak calcium binding affinity that enables it to release or bind calcium in the ER calcium store.
The endoplasmic reticulum (ER) with a resting Ca2+ concentration functions as the primary intracellular Ca2+ store, which can produce both a synchronous Ca2+ release and propagating Ca2+ waves. Ca2+-mobilizing agonists such as ATP, histamine, and glutamine, and second messengers, such as IP3 and cADPR, generate an increase in the cytosolic Ca2+ concentration ([Ca2+]c) with a defined spatio-temporal pattern. The release of Ca2+ from the ER stores results in a rapid increase in [Ca2+]c (from approximately 10−7 M at the resting state to approximately 10−6 M in the excited state) that activates a number of intracellular Ca2+ sensing (trigger) proteins including calmodulin (CaM), troponin C (TnC), and other ion channels and enzymes (Protein Sci. 7: 270-282). While the prevalence of calcium throughout the biological system is well-known and extensive efforts have been made, understanding the calcium regulation of biological functions, stability, folding, and dynamic properties of proteins is limited largely due to the calcium-dependent conformational changes and cooperative calcium binding in natural proteins.
The study of the key determinants of calcium binding has been a continuing endeavor for decades. There are several factors, such as the type, charge, and arrangement of the calcium ligands that have been shown to be important in calcium binding. Calcium is mainly chelated by the oxygen atoms from the sidechains of Asp, Asn, and Glu, the main-chain carbonyl, and solvent water molecules in proteins; the pentagonal bipyramid geometry is the most popular binding geometry. Because of the electrostatic nature of calcium binding, charged Asp and Glu occur most often in calcium binding sites. The charge number in the coordination sphere also plays a role in calcium binding affinity. In addition, a more electronegative environment causes a stronger binding affinity for a given calcium site, and the electrostatic environment affects the cooperativity in multi-site systems. For these multi-site proteins, the apparent calcium affinity contains contributions from the metal-metal interactions and the cooperativity of the binding sites. However, quantitative estimation of the key factors for calcium binding is yet to be established. Therefore, the systematic study of the key determinants for calcium binding required a new strategy and model system.
Monitoring the effects of calcium on the abundant cellular processes has, thus far, been a difficult endeavor due to numerous factors, such as interference from endogenous proteins and perturbation of original calcium signal pathways. While commercially available dyes with binding affinities ranging from 60 nM to hundreds of micromolar can be loaded into mammalian cells through simple incubation, they cannot be targeted to specific cell compartments in a predictable amount, causing difficulty in accurately determining the dye concentration and monitoring calcium concentration. Many of these dyes were shown to have buffering effect in cells and do not provide the necessary sensitivity for thick tissues, intact organisms, or non-mammalian cells. Protein-based calcium sensors that can be directly expressed by the cells and reliably targeted to specific subcompartments have been used in a wide variety of cell types, including mammalian and bacteria. Aequorin was first applied to monitor calcium responses at different cellular environments. However, aequorin requires the constant addition of coelenterazine, which is consumed after each reaction.
FRET-based calcium sensors were then developed using two differently colored fluorescent proteins or their variants linked with a calmdoulin binding peptide and calmodulin (Cell Calcium 22: 209-216; Nature, 388: 882-887). To avoid using the essential trigger protein calmodulin, Troponin C (TnC) was used to sense calcium concentration change in the FRET pair of fluorescent proteins. To address the major concern regarding the competition of endogenous protein and the perturbation of the natural calcium signal systems using essential proteins such as calmodulin and troponin C and the potential perturbation of the natural calcium signal network, a modification of calmodulin binding sites and calmodulin to reduce the interaction was performed (Proc. Natl. Acad. Sci. U.S.A. 101: 17404-17409; Chem. Biol. 13: 521-530). Therefore, there remains a need to develop calcium sensors without using natural calcium binding proteins to monitor the spatial and temporal changes of calcium in the cell, especially at high concentration organelles such as the endoplasmic reticulum.
Endoplasmic reticulum/Sarcoplasmic reticulum calcium signaling are crucial for the research of muscle contraction, brain activity and all the other calcium mishandling related diseases. Different from bulk volume of cytosol in cells, ER/SR has well defined outline and only takes 3% of the total volume of the cell, which is challenged to be studies without highly specific-target calcium indicators. Unfortunately, there are only a few genetically encoded ER calcium sensor published, and all the Kds narrowed around tens of micromolar, while it is well known that free calcium concentration in SR of skeletal muscle cell is around 1 mM, with extra 20 mM calcium bound by calsequestrin. There is a strong need to design an SR calcium sensor with lower binding affinity which is appropriated to measure SR calcium in the muscle cells or tissues. Ideally, the calcium binding affinity should be around 1 mM or sub-millimolar range, similar to the overall calcium binding affinity of SR calcium buffer protein calsequestrin, which is based on the strategy that the cytosolic calcium indicators such as fura-2, camelone and GcamP2 and so on exhibit Kd around sub-micromolar, within the same magnitude of Kd of calmodulin.
The fluorescence change of calmodulin-based calcium sensors highly relies on the interaction between calcium bound form calmodulin and M13 peptide, which is a bulk complex with several different binding processes. The calcium binding affinities to C- and N-domain of calmodulin are in different magnitudes. Moreover, holo-form calmodulin and M13 peptide interact will add an additional Kd to the overall binding process, so the apparent Kd of the sensors does not directly come from the calcium binding, but in a mixture of two Kds with different magnitudes from calcium and calmodulin interaction and a sequential Kd from the calmodulin and M13 peptide interaction. The calmodulin based calcium indicator cannot quantitatively measure the calcium change, as the equation of D1ER binding process involving several constants such as Kd1, Kd2 and Hill coefficients which are difficult to be measured in situ. Furthermore, the kinetics of CaM and M13 peptide interaction could not be further accelerated due to complex delay.